Part 2: Automation to Detect and Stop Fraudulent Transactions

Part 2 Automation to Detect and Stop Fraudulent Transactions Informed

Picking up where we left off last time, welcome to Part 2 of the recap of the BAS Fraud Panel with Informed’s Director of Auto Lending Strategy, Jessica Gonzalez, Kevin Faragher, Senior Director of Product and Strategy at Ally Financial and Michael Reynolds, Business Technology Senior Manager for Service Digitization at KeyBank, moderated by Whitney McDonald, Deputy Editor Of  Bank Automation News. 

During the past week, we’ve seen several more news articles about fraud, so we know this topic is timely and that the panelist’s tips are needed.

Whitney – Kevin and Michael, can you share what you are doing to fight fraud.

Michael – Let’s say you get 10,000 new fraudulent accounts opened… With reporting requirements for local law enforcement and the FBI, you now have 10,000 cases to prepare. And maybe you don’t have staff to do that quickly enough. We know the faster you can react to a situation, the more likely that people will be held accountable . 

From an automation perspective, we look at pulling the necessary data, typically screenshots, reports, slivers of information from a document to support our alleged crime. And then you enter these activities into a certain form. Whether it’s mine or the federal bureau, they have require a certain format. So, from a law perspective, we automate those and prepare the case files. Typically people review and sign off on them and then filter them to the correct agencies.

Whitney – Kevin, tell us what you do.

Kevin – One thing about indirect lending, we always have something we can take back  – like a car.

Interesting enough, you would think that as a lender,  we could turn on OnStar and find our car remotely. But there are privacy rules and regulations, so you can’t. It’s really important. Is it somebody who commits fraud and they actually pay you or do you catch them?

And do you care if you catch them? You probably do because you want the information so you can cycle forward. So a lot of what we do is learning where we can move that into a front end system so when we get those credit applications and contracts from our indirect partners, we can spot them and learn more from the documents.

I read an article that said you can Google free fake fraud documents. There are 40,000 different versions. So there’s no way for humans to deal with that. That is where we get burned. We move information forward and get better about stopping it before it becomes a receivable. 

Whitney – Jessica, do you have anything to add?

Jessica – We make sure that we scour the dark web – Document all of the fake template sites. We look for them consistently, and add to our database. We’re always doing enhancements, making sure we have all of those fraudulent templates and keeping up with changing trends. We also look for missing data that cues fraudulent data within calculations. We have a really low false positive rate – our false positive rate is .02. What that means is that when we tell you that it’s fraud, it is very likely fraud.

We’re not going to say that it is true fraud because we let lenders and banks make those decisions. But you have all the data that’s extracted, so if you want to automate your FCRA process, or other regulatory compliance issues, you can do it and you don’t have to waste valuable man hours. You can state that this is true fraud.

You don’t have to go back and question your fraud team. With informed you can really just say it is fraud, I know it’s fraud, and then start aggregating that data as a true fraud flag. You can look across your portfolio see what other fraud indicators you see, without bias and letting the data speak for itself.

Whitney – Kevin, can you share how Ally handles potential fraud notifications?

Kevin – One situation is a person who commits a crime but actually makes the payments. Sometimes the person who they stole the ID from, if they did steal an ID, goes to get a mortgage and are told they can’t because they have too much debt outstanding. And they say, “I didn’t buy an Escalade,” so they contact us and then we go through a process of ensuring we can help the customer.

But if we discover fraud ourselves then we proactively reach out to the customer who we think has been compromised, let them know, and start the process. You gotta get documentation in place and then you can go into what is most important – helping the person repair their credit history because that is extremely important. 

author avatar
We are digital transformers of tedious manual tasks, bringing robotics process automation to financial services.

New: American Banker - Sharing information = best defense against AI fraud